Hypersaline ecosystems are widespread across the globe, including a wide variety of habitats such as hypersaline lakes, solar salterns, soils, and ancient salt deposits (Oren, 2011). In addition to high salt concentrations, high pH, and low oxygen concentrations are also characteristics of these extreme environments (Naghoni et al., 2017). Although the conditions are harsh, and even high salinity is fatal to most organisms, a large number of halophilic archaea, halophilic and salt-tolerant bacteria still exist in these environments (Oren, 2002; Fendrihan et al., 2006; Naghoni et al., 2017), which play a vital role in global biogeochemical cycles (Mani et al., 2012).

In recent years, the diversity of microorganisms in various hypersaline environments has been studied, but these studies have mainly focused on saline lakes, solar salterns, saline soils (Ben Abdallah et al., 2018; Gomez-Villegas et al., 2018; Bachran et al., 2019; Couto-Rodríguez and Montalvo-Rodríguez, 2019; Nan et al., 2020; Sáenz de Miera et al., 2021). It is worth noting that salt mine, a unique habitat formed tens of millions of years ago, is also an important representative of the hypersaline environment (Chen et al., 2019). For instance, brine from Dingyuan Salt Mine (Dongxing Town, Dingyuan County, Anhui Province, China), an athalassohaline hypersaline environment, is used to produce table salt for hundreds of millions of people. However, current analysis of microbial communities in this kind of environment is limited. Actually, salt mines are rich in microbial resources. For example, the study of halophilic microorganisms in salt mines can not only reveal the evolution and adaptation mechanism of life in extreme environments, but also shows importance in the breeding, development and utilization of microorganisms (Chen et al., 2019).

At present, high-throughput sequencing technology has been applied to investigate microbial communities in different hypersaline environments (Genderjahn et al., 2018; Perez-Fernandez et al., 2019; Zhu et al., 2020). Compared to traditional methods, it has emerged as a reliable tool for investigating differences in microbial community species diversity and structure in any given habitat (Boutaiba et al., 2011), because the culture-independent method of amplicon sequencing appears to be more efficient than the traditional culture-dependent methods (Gibtan et al., 2017). However, culture-dependent method is still a necessary means to acquire valuable microbial strains with potential for new applications and to understand their ecophysiological and environmental functions (Vandamme et al., 1996; Sfanos et al., 2005). For example, the use of traditional culture techniques and the development of new media are encouraged to find novel pure isolates with desired physiological and metabolic characteristics (Rohban et al., 2009).

Therefore, in this study, high-throughput sequencing, clone library and traditional culture-dependent approaches were combined to analyze the microbial diversity of brine (long-term indoor sealed static storage and freshly collected brines) from Dingyuan Salt Mine in Anhui Province (China). The microbial community composition and diversity in this environment were explored using culture-independent and culture-dependent methods, and different results for characterizing microbial communities were compared. In addition, microbial interactions and keystone taxa in complex environments were identified and inferred using co-occurrence network analysis method. The results of this study expand our knowledge on microbial ecology in hypersaline environments.

The Dingyuan Salt Mine in Anhui Province, located in central China, is endowed with the characteristics of continental basin deposits, whose salt-bearing layer was formed in the Early Tertiary Oligocene and is dominated by river-lake facies deposition (Supplementary Figures 1a,b; Chen et al., 2019). At present, water dissolution (fresh water-in and brine-out) is the main method for salt mining (Figures 1A,B). In detail, the mining process involves injecting fresh water (solvent) into the deposit to dissolve salt minerals in situ into a flowing solution (brine), and then the brine coming out (Wang, 1999). In this study, the composition of microbial communities in brines after long-term indoor static and closed storage (C4-5) and in freshly collected brines (C1-3) was preliminarily compared by a combination of culture-independent and culture-dependent methods. Environmental conditions such as high salinity, hypoxia, and slightly alkaline pH are driving factors shaping microbial community structure and forming corresponding adaptability after long-term indoor sealed storage in this study.

It is widely acknowledged that salinity is an important factor affecting the richness and diversity of bacteria and archaea (Kalwasinska et al., 2017; Banda et al., 2021). The DNA concentration of freshwater samples (FW) used to dissolve the rock salt buried underground in this study is significantly higher than that of brine samples (C1-5) (p < 0.05; Supplementary Table 2 and Figure 2A), illustrating that microorganisms in freshwater declined steeply or even disappeared after entering a hypersaline environment. Besides, it also means that environmental filtering prevents the efficient colonization and persistence of non-tolerant species in hypersaline environments (Triado-Margarit et al., 2019). It is worth noting that the bacterial diversity of samples C1-3 is higher than that of archaea, which is in the opposite manner for samples C4-5 (Supplementary Table 3 and Figure 3A). And the diversity of bacterial community in samples C1-3 is significantly higher than that in samples C4-5 (p < 0.05; Supplementary Table 3 and Figure 3A). It is worth noting that some of the error bars in Figures 2A, 3A,B vary greatly, which may be caused by the small sample size. However, the tendency reflected by the differences between these samples (C1-3 and C4-5) are still obvious and persuasive. The result showed that the diversity of bacterial communities in the brine samples decreased significantly after long-term indoor sealed storage. At the phylum level, long-term indoor sealed storage profoundly altered bacterial community structure, i.e., the relative abundance of some bacterial communities varied much more than archaeal communities (Figures 4A,D). At the same time, studies have shown that archaea can withstand environmental stress better than bacteria, and obtain a stable community structure within a certain time-frame (de León-Lorenzana et al., 2017; Mani et al., 2020). Therefore, we believe that archaea in brine samples are more adaptive than bacteria in the hypersaline environment under a long-term indoor sealed storage condition.

In this paper, bacterial and archaeal communities in different brine samples were analyzed by Illumina high-throughput sequencing technology. Overall, the bacterial communities in the brine samples are mainly composed of Proteobacteria and Bacteroidetes (Figure 4A), which is consistent with other researches (Boujelben et al., 2012; Kalwasinska et al., 2018; Mazière et al., 2021). In fact, Proteobacteria and Bacteroidetes are also the dominant phyla in other saline waters (Boujelben et al., 2012; Kalwasinska et al., 2018; Cardoso et al., 2019), playing an important role in carbon and nitrogen cycling (Bernhard et al., 2012; Wong et al., 2015). The result also shows that these brine samples have the characteristics of dominant bacterial taxa commonly found in other hypersaline environments.

Further analysis reveals that although samples C1-3 and C4-5 shared the same dominant bacterial phyla, their relative abundances are quite different (Figure 4A). As it shows in the result that the abundance of Proteobacteria may decrease greatly after long-term indoor sealed storage, whereas the abundance of Bacteroidetes is the opposite. Similarly, genera Salinibacter and Alcanivorax are dominant in C4-5, but less abundant in C1-3 (Figure 4C). Cyanobium_PCC-6307, Aeromonas and Pseudomonas are the dominant genera in C1-3, but are barely detectable in C4-5 (Figure 4C). Among them, genus Salinibacter is an extreme halophilic genus with archaeal properties in Bacteroidetes, also an important part of bacterial communities in various high-salt environments (Bachran et al., 2019; Perez-Fernandez et al., 2019; Cycil et al., 2020). Salinibacter spp. generally, requires light-driven pumps for growth and maintenance of ion gradients across the cell membrane (Doğan and Kocabaş, 2021). The low levels of genus Salinibacter in C1-3 may be attributed to relatively lower levels of solar radiation and lower temperatures (Kalwasinska et al., 2018). Meanwhile, the enrichment of genus Salinibacter is also one of the reasons for the largest proportion of Bacteroidetes in C4-5.

Compared with bacterial taxa, fewer Archaeal phyla can be found in these samples, with most of the sequences belonging to Euryarchaeota, followed by Nanoarchaeaeota (Figure 4D). And the relative abundance difference of the dominant phyla in samples C1-3 and C4-5 is small (Figure 4D). However, the opposite is true at the genus level. The relative abundance of genera Halapricum and Halomicrobium in samples C1-3 is much higher than those in samples C4-5, while the relative abundance of genera Halorientalis, Halobacterium and Haloplanus in samples C4-5 is much higher than those in samples C1-3 (Figure 4F). Interestingly, genus Natronomonas dominates in all samples (Figure 4F). As observed elsewhere, genus Natronomonas is one of the most successful ecological taxa able to survive in hypersaline environments (Chen et al., 2020; Banda et al., 2021; Satari et al., 2021; Zhu et al., 2021).

It should not be ignored that species in genera Alcanivorax and Halorientalis were blooming in hypersaline environment in long-term indoor storage, becoming the dominant genera (Figures 4C,F). Genus Alcanivorax is a ubiquitous marine hydrocarbonoclastic genus, which dominates in many oil-contaminated environments (Scoma and Boon, 2016). Known for its preference for metabolizing hydrocarbons and crude oil derivatives (Zadjelovic et al., 2020). And the members of genus Halorientalis were also reported to be capable of degrading hydrocarbons (Khalid et al., 2021). It is worth noting that the formation process of rock salt and petroleum is always closely related. Petroleum consisting of hydrocarbons with different carbon chain length is covered by a thick stratum of salt. Hydrocarbons with a relatively low density can leak into the upper salt stratum through cracks created by geological movements (Sonnenfeld and Perthuisot, 1984). Additionally, previous studies have shown that the Hefei Basin (Anhui province, China), where the Dingyuan Salt Mine is located, exhibits ideal geological conditions for petroleum formation (Jia et al., 2001; Dai et al., 2011). The brine sample smelled of petrol. Therefore, the brine samples are likely to contain petroleum and hydrocarbons, and such favorable environmental conditions may lead to the enrichment of genera Alcanivorax and Halorientalis. Hydrocarbons degrading microorganisms were enriched in the presence of favorable substrates during a relative long period of storage.

Clone library technology has been widely used to study microbial communities in different habitats (Jones et al., 2009; Xiao et al., 2013; Wang et al., 2018; Chen et al., 2020), which is also applied in this study. It can be found that the dominant genus with the highest proportion in C1-3 is genus Halorientalis, while in C4-5 it is genus Halovenus (Figure 5), which is inconsistent with the high-throughput sequencing results. Previous studies have also shown that the diversity of microbial communities can fluctuate severely along with the size of the clone library (Li et al., 2017). Therefore, it is normal to get different results between the two methods.

With the development of molecular methods, culture-independent methods are considered more effective than culture-dependent methods (Dakal and Arora, 2012), because the latter can only detect 1–5% of all microorganisms in the sample (Ma et al., 2015). Nevertheless, the methods of culture-dependent are still an indispensable technique to acquire microbial species with tremendous application potential and to understand their ecophysiological and environmental functions (Vandamme et al., 1996; Sfanos et al., 2005). In this study, genera Halorubrum, Halobellus, and Halopenitus are the dominant taxa in C4-5-AS168, but genera Natronomonas, Halorubrum, and Haloarcula have a higher proportion in C4-5-NOM (Figure 6B). Among them, the relative abundance of genus Halorubrum in C4-5-AS168 is nearly five times that in C4-5-NOM, indicating that the eutrophic environment of AS-168 may make it more competitive.

Furthermore, a large proportion of microbial species are uncultured, which tends to make some opportunistic species predominant in isolates. Therefore, culture-dependent approaches introduce numerous biases, and generally do not select the most abundant taxa in the environment. Rather, they select the microorganisms that grow best under the culture conditions used. The results often differ from the actual distribution of microbial taxa in the environment (Henriet et al., 2014; Naghoni et al., 2017). In this study, a total of 62 halophilic microorganisms including 12 possible new species (sequence identity < 97.5%) isolated from the two media belong to 10 genera (Figure 6C). Notably, their sequence similarity search based on 16S rRNA gene sequences all show relatively higher sequence identity (>95%). However, most of the 16S rRNA gene sequences obtained by clone library approach exhibit a lower sequence identity (<95%). The unpredictable alteration of template may happen in different cycles during the PCR amplification in clone library construction, which tends to form a larger number of chimeras. Therefore, clone library approach based on PCR amplification of total environmental DNA may severely overestimate the species diversity (Stevens et al., 2013).

Co-occurrence networks can reveal interactions between different taxa in microbial communities, which can be competitive or cooperative (Faust and Raes, 2012; He et al., 2019). In this study, the positive correlation among bacterial networks is 91% (Figure 7A), while that among archaeal networks is only 64% (Figure 7B). This suggests that bacterial communities in brine samples are more likely than archaeal communities to survive in harsh environments through synergies. In addition, genera Ralstonia, Rhodobaculum, uncultured_bacterium_f_Rhodobacteraceae, Salinarimonas, Paracoccus, Cyanobacterium_PCC-10605, uncultured_bacterium_f_Balneolaceae, Algoriphagus and Cyanobium_PCC-6307 are found to be highly associated taxa in the bacterial network. Similarly, genera Haloplanus, uncultured_bacterium_f_Halofracaceae, Halobacterium, Halapricum and Halorientalis are more connected in the archaeal network. These microbial taxa are considered as keystone taxa due to their highly connected nodes (Hu et al., 2018; Guo et al., 2019). Compared with other taxa in the network, these taxa may play a vital part in maintaining the stability of ecological network structure and function (Faust and Raes, 2012; Shi et al., 2019).

Interestingly, the average relative abundances of keystone taxa (except genus Cyanobium_PCC-6307) in the bacterial network were fairly low (0.81%∼2.51%). The results indicated the significance of low-abundance genera in bacterial communities. Although the abundance of such genera may not be high, more attention should be paid to them being the key nodes in bacterial communities (Guo et al., 2019). However, the average relative abundances of keystone genera (except uncultured_bacterium_f_Halofracaceae) in archaeal network ranged from 2.71% to 10.95%. The result of archaea is different from Bacteria in network. The dominant archaea are also key nodes in archaeal network. These differences may be attributed to the different adaptation mechanisms of bacteria and archaea, which are beneficial to their survival and development of their respective taxa. On the other hand, members of the genera Salinarimonas and Paracoccus were previously isolated in oil-contaminated environments, indicating their role in hydrocarbon bioremediation (Zhang et al., 2020; Procópio, 2021). And bacteria belonging to Algoriphagus were also confirmed as an oil-degrading bacterium (Wang et al., 2014). Meanwhile, members of the genera Halobacterium and Halorientalis were reported to be capable of degrading hydrocarbons (Kumar et al., 2020). Therefore, these keystone taxa may play important roles in ecological processes, especially in the remediation of oil-polluted hypersaline environments.

In this work, the microorganisms in the brine of salt mine were studied in details for the first time by combining culture-independent and culture-dependent methods. Our conclusions are:

(1) After long-term indoor airtight storage, the species diversity of bacterial communities in brine samples decreased significantly (p < 0.05), while that of archaeal communities did not change significantly. The composition of the dominant bacterial and archaeal phyla in different samples was similar, but their relative abundances of dominant phyla are significantly different. Among them, halotolerant genera Salinibacter and Natronomonas were the predominant inhabitants in brine samples, suggesting that they play a crucial role in this environment.

(2) A total number of 62 halophilic microorganisms including 12 possible new species (sequence identity < 97.5%) belong to 10 genera were isolated from brine of inland salt mine through culture-dependent method before these inhibiting species went extinct with salt mining. Extremophiles from hypersaline environment are of great significance in special biotechnological applications, and in understanding their ecophysiological and environmental functions.

(3) Network analysis showed that the bacteria in brine samples were more likely than the archaea to survive in harsh environments through synergies. Keystone taxa with highly connected nodes (such as genera Ralstonia, Rhodobaculum, Haloplanus, etc.) play an important role in maintaining the stability of ecological network structure and function.

(4) The brine samples are likely to contain petroleum and hydrocarbons, and such favorable environmental conditions led to the enrichment of specific genera Alcanivorax and Halorientalis, which are capable of degrading hydrocarbons. In addition, the genera Salinarimonas, Paracoccus, Algoriphagus and Halobacterium in keystone taxa also have this ability. The interesting phenomenon reflects that these taxa may have significant contributions in the bioremediation of oil-contaminated hypersaline environments.

Overall, the results of this study will expand the understanding of microbial diversity in extreme environments. On the other hand, it is conducive to explore the functional roles or environmental niches inhabited by various microorganisms in extreme environments.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below (check Supplementary Table 4 for details): NCBI -

(1) PRJNA787045 (https://submit.ncbi.nlm.nih.gov/subs/sra/SUB10779502/overview);

(2) PRJNA791781 (https://submit.ncbi.nlm.nih.gov/subs/sra/SUB10837144/overview);

(3) PRJNA787012 (https://submit.ncbi.nlm.nih.gov/subs/sra/SUB10721350/overview);

(4) PRJNA787052 (https://submit.ncbi.nlm.nih.gov/subs/sra/SUB10779714/overview);

(5) OM184316-OM184500 (https://submit.ncbi.nlm.nih.gov/subs/genbank/SUB10907784/overview);

(6) OL979230-OL979291 (Possible new species: OL979250, OL979251, OL979252, OL979254, OL979260, OL979266, OL979267, OL979268, OL979269, OL979280, OL979282, and OL979283) (https://submit.ncbi.nlm.nih.gov/subs/genbank/SUB10837125/overview).

SC and JH: conceptualization and funding acquisition. SC, SS, DT, JK, YL, and TH: data curation. SC, SS, and DT: investigation. SC and DT: methodology. SC, DT, and JK: writing – original draft. SC, JH, DT, and JK: writing – review and editing. All authors have read and agreed to the published version of the manuscript.